Model-based feedforward-feedback control for real-time hybrid simulation of large-scale structures

Phillips, Brian

Permalink

https://hdl.handle.net/2142/31052 Copy

Description

Title

Model-based feedforward-feedback control for real-time hybrid simulation of large-scale structures

Author(s)

Phillips, Brian

Issue Date

2012-05-22T00:24:36Z

Director of Research (if dissertation) or Advisor (if thesis)

Spencer, Billie F., Jr.

Doctoral Committee Chair(s)

Spencer, Billie F., Jr.

Committee Member(s)

• Elnashai, Amr S.
• Tomizuka, Masayoshi
• Fahnestock, Larry A.

Department of Study

Civil & Environmental Eng

Discipline

Civil Engineering

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Structural Dynamics
• Hybrid Simulation
• Real-Time Hybrid Simulation
• Actuator Control
• Shake Table Testing

Abstract

Extreme dynamic events continue to demonstrate the fragility of civil infrastructure worldwide. Over the years, design codes and computational tools have come to reflect an improved understanding of dynamic loads and effects. However, experimental testing often drives these changes. Experimental testing is vital to understanding the behavior of structures subjected to these dynamic loads and evaluating new solutions for hazard mitigation. Common experimental frameworks include quasi-static testing, shake table testing, and hybrid simulation. The tradeoffs in loading protocol make each experimental framework attractive in different situations. Hybrid simulation is a powerful, cost-effective framework for testing structural systems, closely coupling numerical simulation and experimental testing to obtain the complete response of a structure. Through substructuring, the well-understood components of the structure are modeled numerically, while the components of interest are tested physically. Generally, an arbitrary amount of time may be used to calculate and apply displacements at each step of the hybrid simulation. However, when the rate-dependent behavior of the physical specimen is important, real-time hybrid simulation (RTHS) must be employed. In RTHS, computation, communication, and actuator limitations cause delays and lags which lead to inaccuracies and potential instabilities. At the same time, the phenomenon of control-structure interaction (CSI) leads to a coupling of the dynamic behavior of the actuators and the structure. Traditional actuator control approaches for RTHS compensate for an apparent time delay or time lag rather than address the actuator dynamics directly. Furthermore, most actuator control approaches focus on single-actuator systems. The model-based actuator control approach proposed herein directly addresses actuator dynamics including CSI and actuator coupling through model-based feedforward-feedback control. The feedback controller is flexible to include multi-metric measurements for improved tracking of higher-order derivatives, moving beyond the traditional focus solely on displacement tracking. The proposed approach is illustrated for predefined trajectories as well as RTHS of both single and multi-actuator systems. The similarities between actuator control for RTHS and shake tables were leveraged to apply the proposed model-based actuator control approach to acceleration tracking. Shake tables provide a direct means by which to evaluate structural performance under earthquake excitation. Essentially, an actuator excites a base plate on which a structural model is mounted with a predefined acceleration record. Improved acceleration tracking will be explored in the presence of large nonlinearities in shake table behavior as well as changes in shake table dynamics through CSI. The research presented in this dissertation provides an advanced framework for the dynamic performance evaluation of structural systems. A broad class of structures is considered for RTHS, including multi-degree-of-freedom (MDOF) structures through accurate control across a broad frequency range, multi-actuator systems through the modeling of actuator coupling, and improved tracking of higher-order derivatives through multi-metric feedback control. Application to shake table control demonstrates the versatility of the proposed actuator control scheme for general real-time actuator control.

Graduation Semester

2012-05

Permalink

http://hdl.handle.net/2142/31052

Copyright and License Information

© 2012 Brian M. Phillips

Owning Collections